After building a few AI apps with LLMs and LangChain (including https://github.com/orgexyz/BlockAGI), we finally felt that it's time for a change.

LLFn was born out of our necessity to iterate faster on building AI agents. We found that the current frameworks are convoluted with complex abstractions which over time has present more of a road block than saving our time.

LLFn is an open-source framework that turns any prompt into a function and let you decoratively define their output. This abstraction allows us to build AI apps, especially those with multiple agents interacting with each other, much faster. Because each "agent" is just a function, you can compose it any way you want with minimum boilerplate. This turned out to be 10x productivity for our team.

We're super excited to share it with you and hear your feedback!

https://github.com/orgexyz/LLFn

If you want to try LLFn on Colab, here's the link:

https://colab.research.google.com/drive/11cYRchBaKftxGJvyacd...